校准SMA JACK-IPEX MHFI 95MM适配器的最佳方法是什么？

有一种可能的方法可以做到这一点，它被称为单端口AFR，现在可用于PLTS和PNA，但不能用于ENA。
如果您可以校准到SMA连接器并测量10 MHz至20 GHz的开放式适配器的s1p，AFR可以从s1p文件中提取2端口适配器。
然后，您可以校准SMA并使用适配器测量设备，使用去嵌入从DUT测量中移除适配器。
你会附上你的连接图片，以确保我们彼此了解，并重新测量从10MHz到20GHz的开放式适配器？
然后我将尝试为您提取2端口适配器文件。



     以下为原文

  There is one possible way to do this and it is called one-port AFR,available in PLTS and PNA now,but not in ENA. If you can calibrate to the SMA connector and measure the s1p of the open-ended adapter from 10 MHz to 20 GHz,AFR can extract the 2-port adapter from the s1p file. Then you can calibrate to the SMA and measure your device with the adapter,use de-embeding to remove the adapter from your DUT measurement. 
Would you attach a picture of your connections to make sure we understand each other, and re-measure your open-ended adapter from 10MHz to 20 GHz? Then I will try to extract the 2-port adapter file for you.

> {quote：title = Sharpie写道：} {quote}>嗨我正在使用E5017C VNA和相应的ECal模块。
要访问我的DUT，我必须使用SMA-U.FL连接器，我不能将其包含在ECal校准中。
校准SMA JACK-IPEX MHFI 95MM适配器的最佳方法是什么？
我正在附加SMA JACK的S1P文件 - IPEX MHFI 95MM电缆，U.FL端打开。
>>谢谢ningchen很友好地提出要为你做这件事（我假设），把它加载到PLTS或PNA中。
让他接受这个提议是有意义的。
我不知道该方法的假设是什么 - 例如假设理想化开放，还是尝试考虑边缘电容？
但是如果你想要另一种可行的方式，但我认为精度较低，你可以使用端口扩展。
我会在短时间内终止电缆的末端，然后调整端口延伸直到相位为180度。
我已经用UFL连接器完成了几千GHz，并取得了相当大的成功。
端口扩展避免了你购买另一个VNA或一些10,000美元的软件，但我怀疑它与自动夹具移除一样准确，虽然我不知道它有多聪明。
戴夫



     以下为原文

  > {quote:title=Sharpie wrote:}{quote}
> Hi I am using an E5017C VNA with the corresponding ECal module. To access my DUT I have to use an SMA-U.FL connector which I cannot include in the ECal calibration. What is the best way to calibrate out the presence of the SMA JACK - IPEX MHFI 95MM adapter? I am attaching the S1P file of the SMA JACK - IPEX MHFI 95MM cable with the U.FL end open.
> 
> Thank you

ningchen had kindly offered to do this for you by (I assume), loading it into PLTS or an PNA. It makes sense to take him up on that offer. 

I don't know what assumptions that method makes - does it for example assume an idealised open, or does it attempt to take into account the fringing capacitance? 

But if you want another way, which will work, but I assume at less accuracy, you can use a port extension. I would terminate the end of the cable in a short, and then adjust the port extension until the phase is 180 degrees. I've done that up to a few GHz with UFL connectors and had a reasonable amount of success. 

The port extension avoids you buying another VNA or a bit of $10,000 software, but I doubt it it as accurate as the automatic fixture removal, although I don't know how clever that is. 

Dave

> {quote：title = drkirkby写道：} {quote}>端口扩展避免你购买另一个VNA或一些10,000美元的软件，但我怀疑它是否像自动夹具一样准确，虽然我不知道多么聪明
那是。
>> Dave AFR代码支持两种模式：仅2端口或1端口AFR，仅打开或仅打开。
两个版本通过计算适配器的完整s2p文件完全纠正所有适配器的错误，包括输入匹配，输出匹配和S21 mag和相位（我们假设S21 = S12）。
它依赖于时域变换和“其他数据处理”的组合来获得这些值。
它还要求固定装置在DUT接口附近具有恒定的imedance（不一定是50欧姆但是恒定）。
这里的“vincinity”意味着没有一个距离代表分析仪最高频率的4个上升时间。
对于26.5 GHz分析仪，如果我正确计算，这相当于大约2.25厘米（小于1英寸）。
也就是说，固定装置应该在靠近dut的恒定阻抗线上长约1英寸。
如果使用50 GHz分析仪，则可将其推至约1 cm。
您可以在夹具上使用更高的频率测量值一次来计算AFR结果，然后获取S2P文件并在较低频率的分析仪上使用它进行持续测试。



     以下为原文

  > {quote:title=drkirkby wrote:}{quote}

> The port extension avoids you buying another VNA or a bit of $10,000 software, but I doubt it it as accurate as the automatic fixture removal, although I don't know how clever that is. 
> 
> Dave

The AFR code supports two modes: 2-port thru only, or 1 port AFR with either open only or short only. Both versions fully correct for all the adapter's errors by computing the full s2p file for the adapter including input match, output match and S21 mag and phase (we presume S21=S12).  It relies on a combination of time-domain transforms and "other data processing" to obtain these values.  It also requires that the fixture be of constant imedance (not necessarily 50 ohms but constant) in the vicinity of the DUT interface.  Here "vincinity" means without about a distance that represents 4 risetimes of the highest frequency of the analyzer.  For a 26.5 GHz analyzer, this amounts to about 2.25 cm (less than 1 inch) if I figure it correctly.  That is, the fixture should have a constant impedance line near the dut that is about 1 inch long. If you use a 50 GHz analyzer, you can push that down to about 1 cm.  You can use a higher freq measuerment on a fixture one time to compute the AFR result, then take the S2P file and use it on a lower frequency analyzer for ongoing test.

> {quote：title = Dr_joel写道：} {quote} >> {quote：title = drkirkby写道：} {quote} >>>端口扩展可以避免你购买另一个VNA或一些10,000美元的软件，但我对此表示怀疑
和自动夹具拆卸一样准确，虽然我不知道它有多聪明。
>>>> Dave >> AFR代码支持两种模式：仅2端口或1端口AFR，仅打开或短路。
两个版本通过计算适配器的完整s2p文件完全纠正所有适配器的错误，包括输入匹配，输出匹配和S21 mag和相位（我们假设S21 = S12）。
显然比我使用的端口扩展要好很多！
>它依赖于我认为会有时域变换的组合 - 你的书中有一些关于这类事情的信息。
>和“其他数据处理”以获得这些值。
魔术！
>它还要求灯具在DUT接口附近具有恒定阻抗（不一定是50欧姆但是恒定）。
这里“附近”表示没有表示分析仪最高频率的4个上升时间的距离。
对于26.5 GHz分析仪，如果我正确计算，这相当于大约2.25厘米（小于1英寸）。
也就是说，固定装置应该在靠近dut的恒定阻抗线上长约1英寸。
如果使用50 GHz分析仪，则可将其推至约1 cm。
您可以在夹具上使用更高的频率测量一次来计算AFR结果，然后获取S2P文件并在较低频率的分析仪上使用它进行持续测试。
如果我错了，请纠正我，但是导出的S参数AFR不一定是参考平面，这可能导致重大错误。
我在这个例子中使用了一个“N”连接器，原因我猜对你很明显，但原理是一样的。
假设有一个26.5 GHz 3.5 mm校准套件而不是N cal套件，但是想要测试一个N型设备。
可以用3.5 mm进行校准，然后使用一端为3.5 mm的适配器，另一端使用N型。
如果对适配器进行单端口测量，则使用自动夹具移除，我认为会出现严重错误，因为阴N的引脚末端距参考平面约5 mm - 外导体配合平面。
我对么？
（同样令人怀疑的是，适配器能够满足高达18 GHz的恒定阻抗条件，但这是另一个问题，但对于UFL适配器也是如此。这些都是几个$ s，所以不完全是精密设备。
）我不认为这对于U.FL的SMA是一个问题，因为连接器的参考平面几乎与母针的末端处于相同的位置。
我有兴趣知道是否可以使用AFR来测量BNC DUT，使用1）20 GHz VNA 2）85054B 18 GHz N cal套件3）N到BNC适配器作为夹具的组合。
4）N1930B物理层测试系统（PLTS）在单端口模式下，BNC连接器刚刚打开。
我拥有1-3，并且知道有PLTS的人。
如果没有BNC连接器在我面前，我不确定引脚相对于参考平面的位置，但我知道N，母中心导体的末端远离参考平面。
戴夫



     以下为原文

  > {quote:title=Dr_joel wrote:}{quote}
> > {quote:title=drkirkby wrote:}{quote}
> 
> > The port extension avoids you buying another VNA or a bit of $10,000 software, but I doubt it it as accurate as the automatic fixture removal, although I don't know how clever that is. 
> > 
> > Dave
> 
> The AFR code supports two modes: 2-port thru only, or 1 port AFR with either open only or short only. Both versions fully correct for all the adapter's errors by computing the full s2p file for the adapter including input match, output match and S21 mag and phase (we presume S21=S12).  

Obviously a lot better than the port extension I have used! 

> It relies on a combination of time-domain transforms 

I thought it would - there is some info in your book about this sort of thing. 

> and "other data processing" to obtain these values. 

The magic! 

>  It also requires that the fixture be of constant impedance (not necessarily 50 ohms but constant) in the vicinity of the DUT interface.  Here "vicinity" means without about a distance that represents 4 risetimes of the highest frequency of the analyzer.  For a 26.5 GHz analyzer, this amounts to about 2.25 cm (less than 1 inch) if I figure it correctly.  That is, the fixture should have a constant impedance line near the dut that is about 1 inch long. If you use a 50 GHz analyzer, you can push that down to about 1 cm.  You can use a higher freq measurement on a fixture one time to compute the AFR result, then take the S2P file and use it on a lower frequency analyzer for ongoing test.

Correct me if I am wrong, but the S-parameters AFR derives are not necessarily to the reference plane, which could lead to significant errors. I use an "N" connector in this example for reasons I guess are obvious to you, but the principle is the same. 

Lets assume  one had a 26.5 GHz 3.5 mm cal kit and not an N cal kit, but wanted to test a male N device. One could calibrate with 3.5 mm, then use an adapter with 3.5 mm one end, and female N on the other. If one does a one port measurement of the adapter, then used the automatic fixture removal, I think there would be serious errors as the end of the pin of the female N is about 5 mm away from the reference plane - outer conductor mating plane. 

Am I correct? (It's also doubtful the adapter would meet the condition of constant impedance up to 18 GHz, but that is another issue, but the same could be said for the UFL adapter. These are a few $s each, so not exactly a precision device. )

I don't see this being an issue with an SMA to U.FL, as the reference plane of the connector is pretty much in the same position as the end of the female pin. 

I'd be interested to know if one could use the AFR to measure a BNC DUT using a combination of 

1) 20 GHz VNA
2) 85054B 18 GHz N cal kit
3) N to BNC adapter as a fixture. 
4) N1930B Physical Layer Test System (PLTS) in one-port mode with the BNC connector just open. 

I own 1-3, and know someone with PLTS. 

Without a BNC connector in front of me, I'm not exactly sure where the pins are relative to the reference plane, but I know for an N, the end of the female centre conductor is nowhere near the reference plane. 

Dave